Accumulating evidences indicate that long non-coding RNAs (lncRNAs) play pivotal roles in various biological processes. Mutations and dysregulations of lncRNAs are implicated miscellaneous human diseases. Predicting lncRNA–disease associations is beneficial to disease diagnosis as well treatment. Although many computational methods have been developed, precisely identifying associations, especially for novel lncRNAs, remains challenging. In this study, we propose a method (named SIMCLDA)...

Abstract Motivation Computational drug repositioning is an important and efficient approach towards identifying novel treatments for diseases in discovery. The emergence of large-scale, heterogeneous biological biomedical datasets has provided unprecedented opportunity developing computational methods. problem can be modeled as a recommendation system that recommends based on known drug–disease associations. formulation under this matrix completion, assuming the hidden factors contributing...

Abstract Motivation Protein–protein interactions (PPIs) play important roles in many biological processes. Conventional experiments for identifying PPI sites are costly and time-consuming. Thus, computational approaches have been proposed to predict sites. Existing methods usually use local contextual features Actually, global of protein sequences critical site prediction. Results A new end-to-end deep learning framework, named DeepPPISP, through combining sequence features, is For we a...

Abstract Motivation Computational drug repositioning is a cost-effective strategy to identify novel indications for existing drugs. Drug often modeled as recommendation system problem. Taking advantage of the known drug–disease associations, objective new treatments by filling out unknown entries in association matrix, which matrix completion. Underpinned fact that common molecular pathways contribute many different diseases, assumes underlying latent factors determining associations are...

Abstract Biomolecular recognition between ligand and protein plays an essential role in drug discovery development. However, it is extremely time resource consuming to determine the protein–ligand binding affinity by experiments. At present, many computational methods have been proposed predict affinity, most of which usually require 3D structures that are not often available. Therefore, new can fully take advantage sequence-level features greatly needed accelerate process. We developed a...

In recent years, the recommendation systems have become increasingly popular and been used in a broad variety of applications. Here, we investigate matrix completion techniques for that are based on collaborative filtering. The filtering problem can be viewed as predicting favorability user with respect to new items commodities. When rating is constructed users rows, columns, entries ratings, then modeled by filling out unknown elements matrix. This article presents comprehensive survey...

ICD-9 (the Ninth Revision of International Classification Diseases) is widely used to describe a patient's diagnosis. Accurate automated coding important because manual expensive, time-consuming, and inefficient. Inspired by the recent successes deep learning, in this study, we present learning framework called DeepLabeler automatically assign codes. combines convolutional neural network with `Document Vector' technique extract encode local global features. Our proposed demonstrates its...

The explosion of digital healthcare data has led to a surge data-driven medical research based on machine learning. In recent years, as powerful technique for big data, deep learning gained central position in circles its great advantages feature representation and pattern recognition. This article presents comprehensive overview studies that employ methods deal with clinical data. Firstly, the analysis characteristics various types (e.g., images, notes, lab results, vital signs demographic...

High-throughput screening technologies have provided a large amount of drug sensitivity data for panel cancer cell lines and hundreds compounds. Computational approaches to analyzing these can benefit anticancer therapeutics by identifying molecular genomic determinants developing new drugs. In this study, we developed deep learning architecture improve the performance prediction based on data. We integrated both features chemical information compounds predict half maximal inhibitory...

Purpose – Social media analytics uses data mining platforms, tools and techniques to collect, monitor analyze massive amounts of social extract useful patterns, gain insight into market requirements enhance business intelligence. The purpose this paper is propose a framework for competitive intelligence value Design/methodology/approach authors conducted case study collect set with nearly half million tweets related two largest retail chains in the world: Walmart Costco past three months...

Computational methods including centrality and machine learning-based have been proposed to identify essential proteins for understanding the minimum requirements of survival evolution a cell. In methods, researchers are required design score function which is based on prior knowledge, yet usually not sufficient capture complexity biological information. some selected features cannot represent complete properties information as they lack computational framework automatically select features....

Abstract Annotation of protein functions plays an important role in understanding life at the molecular level. High‐throughput sequencing produces massive numbers raw proteins sequences and only about 1% them have been manually annotated with functions. Experimental annotations are expensive, time‐consuming do not keep up rapid growth sequence numbers. This motivates development computational approaches that predict A novel deep learning framework, DeepFunc, is proposed which accurately...

The purpose of this paper is to identify at-risk online students earlier, more often, and with greater accuracy using time-series clustering. case study showed that the proposed approach could generate models higher feasibility than traditional frequency aggregation approaches. best performing model can start capture from week 10. In addition, four phases in student's learning process detected holiday effect illustrate students' behaviors before after a long break. findings also enable...

In bioinformatics, machine learning-based prediction of drug-target interaction (DTI) plays an important role in virtual screening drug discovery. DTI prediction, which have been treated as a binary classification problem, depends on the concentration two molecules, between and other factors. The degree affinity molecule (such compound) target receptor or protein kinase) reflects how tightly binds to particular is quantified by measurement can reflect more detailed specific information than...

Identifying essential proteins plays an important role in disease study, drug design, and understanding the minimal requirement for cellular life. Computational methods discovery overcome disadvantages of biological experimental that are often time-consuming, expensive, inefficient. The topological features protein-protein interaction (PPI) networks used to design computational prediction methods, such as Degree Centrality (DC), Betweenness (BC), Closeness (CC), Subgraph (SC), Eigenvector...

Abstract With the development of high-throughput technology and accumulation biomedical data, prior information biological entity can be calculated from different aspects. Specifically, drug–drug similarities measured target profiles, interaction side effects. Similarly, methods data sources to calculate disease ontology result in multiple measures pairwise similarities. Therefore, computational drug repositioning, developing a dynamic method optimize fusion process is crucial challenging...

We apply generative adversarial network (GAN) technology to build an event generator that simulates particle production in electron-proton scattering is free of theoretical assumptions about underlying dynamics. The difficulty efficiently training a GAN simulator lies learning the complicated patterns distributions particles physical properties. develop selects set transformed features from momenta can be generated easily by generator, and uses these produce augmented improve sensitivity...

Abstract Motivation The identification of compound–protein interactions (CPIs) is an essential step in the process drug discovery. experimental determination CPIs known for a large amount funds and time it consumes. Computational model has therefore become promising efficient alternative predicting novel between compounds proteins on scale. Most supervised machine learning prediction models are approached as binary classification problem, which aim to predict whether there interaction...

The identification of drug–target relations (DTRs) is substantial in drug development. A large number methods treat DTRs as drug-target interactions (DTIs), a binary classification problem. main drawback these are the lack reliable negative samples and absence many important aspects DTR, including their dose dependence quantitative affinities. With increasing publications drug–protein binding affinity data recently, prediction can be viewed regression problem affinities (DTAs) which reflects...

Randomized algorithms for low-rank matrix approximation are investigated, with the emphasis on fixed-precision problem and computational efficiency handling large matrices. The based so-called QB factorization, where Q is an orthonormal matrix. First, a mechanism calculating error in Frobenius norm proposed, which enables efficient adaptive rank determination and/or sparse It can be combined any QB-form factorization algorithm B's rows incrementally generated. Based blocked randQB by...

Modeling loops is a critical and challenging step in protein modeling prediction. We have developed quick online service (http://rcd.chaconlab.org) for ab initio loop combining coarse-grained conformational search with full-atom refinement. Our original Random Coordinate Descent (RCD) closure algorithm has been greatly improved to enrich the sampling distribution towards near-native conformations. These improvements include new workflow optimization, MPI-parallelization fast backbone angle...